Anthropomorphic reference track planning method for unmanned driving

Abstract

The invention relates to an anthropomorphic reference track planning method for unmanned driving, including the following steps: building a vehicle steering model, and generating dynamic constraints of vehicle steering according to the basic parameters of vehicles; acquiring map information of roads according to a vehicle top decision, and getting start and end point constraint information of a reference track according to the map information; according to the dynamic constraints of vehicle steering and the start and end point constraint information of the reference track, generating a plurality of turning track curves which basically meet the geometric and dynamic constraints of vehicle steering and accord with people's driving habits and steering law, and forming a basic feasible track cluster; and with the integral of lateral acceleration over time as an index, selecting a most comfortable track as a finally generated optimal reference driving track within the intersection from the basic feasible track cluster. As the driving characteristic of people is considered and the driving law of people is simulated, the reference driving track generated is in accordance with the driving habits of people.

Description

Technical field [0001] The invention belongs to the field of unmanned car navigation, and in particular relates to an unmanned driving-oriented anthropomorphic reference trajectory planning method. Background technique [0002] The unmanned vehicle navigation module integrates the results of perception and considers the dynamics of the vehicle to output a trajectory that can be driven. In the fully autonomous car navigation of urban structured roads, navigation tasks are generally divided into three layers: task planning layer, behavior decision layer and trajectory planning layer. The trajectory planning layer is generally divided into two steps: reference trajectory generation without considering moving obstacles and local trajectory planning considering moving obstacles. For the problem of trajectory planning, the reference trajectory is particularly important in complex scenes such as intersections. It compensates for the incompleteness of upper-level mission planning, while...

AI Technical Summary

Problems solved by technology

[0003] At present, the generation method of the turning reference trajectory at the intersection is not perfect. Generally, it is generated by straight line connection, arc connection or manual drawing of virtual lane lines on the map. Factors have not been considered in depth

In addition, the manual drawing method itself is not efficient and is not suitable for large-scale driving navigation of unmanned vehicles

Images